Minutes for the OGSA Data WG telcon, December 21st 2005

1) Early discussion
 
* Roll call 

Dave Berry, NeSC
Allen Luniewski, IBM
Ann Chervenak, ISI
Chris Jordan, SDSC

Minutes were approved.

2) Action report

* New actions

Allen to work on an integration pass, including a scoping section.  Target c.o.b. Friday 6th January.
Dave to iterate the description of the ByteIO interface (and let Mario know).

* Completed actions

- Mark to post UVa's "RNS-lite" document, requesting people to comment. [Done]
- Ann to update replication section.  [Done]
- Dave to revamp section 3 of the architecture document [Done]
- Allen & Dave to mail GFS co-chairs saying that we want to collaborate on architectural design. [Done]

* Ongoing actions

- Dave to investigate Skype or phone for OGSA Data session at F2F. [Mailed Hiro, haven't heard back].
- Stephen to investigate whether transient data is supported by RNS.
- Allen & Dave to draft message to GFS co-chair & ADs, saying RNS is too complicated. [Done.  Dave & Allen to send to Ann & Chris for review]

- Ann to check the Access section for suitability for file replication.
- Mario to update the Access section and make text available for comment. [Partially done]
- Section authors to add security notes to their sections.
- Everyone to talk with your contacts to get more participation in this WG.
- Allen/Dave to contact David Martin & Hiro Kishimoto to get reviewers & expert participation (after internal reviews).  


3) Progess & planning update

Chris planning to attend Sunnyvale on 16th/17th January.
We agreed the proposed revision of session times for 19th January.


4) Review of architecture document

Allen: The document needs an integration pass.

Dave suggested a scoping subsection (in section 2?) saying we're working on files & databases but not streams.  Then delete the placeholder for section 15.  Also have a section at the end on "other services", including personal data service, provenance services, transformation services.

Chris raised a concern about the scope of transformational services.  Does this include major computation?  Allen thinks that transformation should apply only to data in motion.  So perhaps it should apply only within data transfer.  We also talk about transformation in the data federation section.  (NOte: we use the term "data treatment" in some parts of the document.  Ann used the term "data mediation" for schema translation).  Decision: don't have a section on data transformation per se, but just describes appropriate transformations where they apply in the individual sections.

Dave: how to describe interfaces?  We want to avoid details because we don't want to track changes.  Perhaps textual descriptions of the key operations and properties?

Allen: the key operations are those that help to tie the components together.  This is a fuzzy concept.  The best way forward might be to iterate a couple of descriptions until we're happy and then use those as exemplars.  Dave to attempt an iteration of the ByteIO example (and let Mario know I'm doing that).

Ann: some APIs don't exist yet, so we can't give them in detail anyway.  In these cases we can say that an interface SHOULD provide certain key operations.

Allen to work on an integration pass, including a scoping section.  Target c.o.b. Friday 6th January.


5) Wrap up

Next meetings:

Dec 28th:	No meeting.
Jan 4th:	No meeting.
Jan 11th:   (Planning for F2F, Status review of documents, ...)
Jan 16th-23rd:	OGSA F2F in Sunnyvale